Physical Database DDL Statements › AREA Statements › Entity Currency › Usage

Usage

Unique Page Range

The range of pages reserved for an area is defined by the FROM PAGE parameter in conjunction with the MAXIMUM SPACE parameter (or the PRIMARY SPACE parameter if you do not specify MAXIMUM SPACE). This page range must not overlap the page range for:

• Any other area contained in the segment
• Any other area in a DMCL in which the area's segment is included if the page groups are the same

Contiguity of Page Ranges

Page ranges within a segment can be, but do not have to be, contiguous with one another.

Page Range Limits Depend on Maximum Number of Records Per Page

The highest page number for an area depends on the maximum number of records or rows that can fit on a single page. Use the table provided under "Usage" in 7.16, “SEGMENT Statements” to determine the highest page number.

Page Ranges for CALC Records

The last page of a subarea that can be used to store CALC record occurrences depends on the type of offset specification:

• For page offsets, the last page of the CALC range is the last page of the subarea.
• For percentage offsets, this calculation is used to determine the last page of the CALC range:

  calc-lastpage-of-subarea =
      firstpage-of-subarea + percent * primary-page-count * .01
  

What Happens to Offsets When You Expand an Area

When you expand an area by using the EXTEND SPACE clause on the ALTER AREA command, the following occurs to the first page, last page, and CALC last page of a subarea:

• The first page does not change
• The last page changes if you specified a percentage offset; CA IDMS/DB allows CALC overflow records and records with other location modes to be stored in the expanded space
• The last page of the CALC range does not change; that is, CALC records continue to target to the original page range assigned to the subarea

Note: You must exercise care when expanding an area containing subarea definitions that use offset percentages because subareas can overlap after the EXTEND SPACE is performed. For example:

Given the following subarea allocations for an area containg 1000 pages, the page ranges are as follows:

    Definition                LoPage     HiPage
SUB1 OFFSET  0% FOR 25%          1         250
SUB2 OFFSET 25% FOR 25%        251         500
SUB3 OFFSET 50% FOR 25%        501         750
SUB4 OFFSET 75% FOR 25%        751        1000

If an EXTEND SPACE is executed, and 1000 more pages are added to the area, the allocations are as follows:

Sub Area        LoPage        HiPage
  SUB1             1            500
  SUB2           251            750
  SUB3           501           1000
  SUB4           751           1250

You can see that adding 1000 pages to the area did not significantly increase the space available in which to store records, nor did the additional space get used by any areas mapped to defined subarea definitions—only SUB4 benefits by the additional pages.

Percentage Offsets Most Flexible

Percentage offset specifications are the most flexible in terms of database maintenance. As the database grows and must eventually be expanded, the areas of the database must also be expanded. If you use percentage offsets, CA IDMS/DB automatically assigns record occurrences to the appropriate percentage of the new area.

Page Range for RRDS Native VSAM Areas

CA IDMS/DB constructs the db-key for a record in an RRDS native VSAM area in the following manner:

dbkey = low-dbkey-of-area + relative-record-number

Therefore, for an RRDS file, the number of pages specified by the page range must be calculated as follows (rounded up to the next integer):

number-of-pages =
 (number-of-vsam-records-in-file + 1) / (maximum-records-per-page + 1)

Note: Maximum-records-per-page is specified on the CREATE SEGMENT statement and determines the format of the database keys for records in areas that are contained in the segment.

Page Range for RRDS Native VSAM Areas

CA IDMS/DB constructs the db-key for a record in a KSDS native VSAM area by randomizing the record's prime key to a database key in the database key range for the area. Therefore, for a KSDS file, a rule-of-thumb for calculating the page range is as follows (rounded up to the next integer):

number-of-pages = number-of-vsam-records-in-file / x

where x =  10 if number-of-vsam-records-in-file < 100,000
           100 if number-of-vsam-records-in-file > 100,000

The idea is to specify a page range that minimizes the probability of constructing duplicate keys without specifying an excessive number of pages for the area.

Page Range for ESDS Native VSAM Areas

CA IDMS/DB constructs the db-key for a record in an ESDS native VSAM area in the following manner:

dbkey = low-dbkey-of-area + relative-byte-address

Therefore, for an ESDS file, the number of pages specified by the page range must be calculated as follows (rounded up to the next integer):

number-of-pages = total-bytes-in-file / (maximum-records-per-page + 1)

Note: Maximum-records-per-page is specified on the CREATE SEGMENT statement and determines the format of the database keys for records in areas that are contained in the segment.

Physical Device Blocking

A database page is a fixed block. As a general rule, you should use pages that are an even fraction of the track size.

The following table lists the optimal page sizes by device type for five IBM disk drives. Manufacturers of other brands of direct access storage devices (DASD) should be able to provide similar information for their own equipment.

Note: The bytes per page for FBA devices must be a multiple of 512.

Note: On z/VM, the size of a database page must be less than or equal to 4096 bytes.

Note: For VSAM database files the character-count must be at least 8 bytes larger than the page size.

Synchronization Stamps

If you expect frequent changes to the definitions of SQL tables, you should maintain synchronization stamps at the table level. If you do not expect frequent changes, you should maintain stamps at the area level.

Specifying a Synchronization Stamp Value

When defining or altering an area for which area-level stamping is in effect, you can specify an explicit value for its synchronization stamp. This allows you to create databases that have identical physical attributes and can therefore be accessed through a single schema definition.

Since an area's synchronization stamp is updated each time any DDL statement affecting the area is issued, the synchronization stamp must be set after issuing the SQL DDL statements that define the database.

Care should be exercised when specifying a specific timestamp, since its purpose is to enable the detection of discrepancies between an entity and its definition. If explicitly specified, the timestamp should always be set to a new value following a definitional change so that the change is detectable to the run time system.

Contiguity of File Blocks

Block ranges within a file associated with more than one area must be contiguous.

To specify that all pages of the area map to all pages of the file, specify:

...from 1 for all

on the file specification.

If the file has multiple areas associated with it, the block range will overlap if both of the areas map to the file having this specification. You can map the first area using "FROM 1 FOR ALL", but you must map the second area "FROM last-block-of-the-file+1 FOR ALL".

Device Types or Access Methods May Limit the Number of File Blocks

Device types or access methods may further restrict the number of blocks allowed in a file. For example, a maximum of 65,535 tracks can be addressed in BDAM files.

Note: For more information about device types and access methods, see Allocating and Formatting Files.

Native VSAM File Restrictions

An area that maps to native VSAM files has the following restrictions:

• A native VSAM file defined with an access method RRDS, KSDS, or ESDS can map to one and only one area
• An area that maps to a native VSAM file must map to one and only one file
• If an area is associated with one or more files defined with PATH as an access method, then:
  • The area must map to either an ESDS or KSDS file
  • The PATH file must not be associated with any other area

Index Calculations

The following algorithms are used to calculate BLOCK CONTAINS key-count and the DISPLACEMENT page-count values for symbolic index parameters when the BASED ON clause is specified.

Index block:

Step 1:  Assuming 3 SR8's per page, compute the following:

         The maximum size of the variable portion of an SR8:
            ( (Page size - Page reserve - 32) / 3 ) - 40 = SR8-vsize

         The maximum number of entries in an SR8:
                Sorted index:  (SR8-vsize / ( 8 + Keylen) ) - 2
              Unsorted index:  (SR8-vsize / 4 ) - 1

         If the number of SR8 entries is less than 3, set it to 3; if
         greater than 8180, set it to 8180.

Step 2:  Establish the number of index entries:  Use the FOR index-cnt
         value, if specified, or 1000.

Step 3:  Estimate the number of entries per SR8 for a 3-level index:
         Find the first entry in the following table whose Number of
         Entries column is greater than or equal to the value established
         in Step 2.

               Number of        Number of
               entries          SR8 entries

                   1,000           10
                  15,625           25
                 125,000           50
                 512,000           80
               1,000,000          100
               2,000,376          126
               3,375,000          150
               5,359,375          175
               8,000,000          200
              15,625,000          250
                 -1              8180

Step 4:   Determine the INDEX BLOCK value:  Use the lesser of the Number
          of SR8 entries from the table and the value from Step 1 as the
          INDEX BLOCK (IBC) value in the remaining calculations.

Displacement:

For unsorted indexes, the displacement is set to 0; for sorted indexes, it is calculated as follows:

Step 1:  Calculate the number of bottom level and higher level SR8s:

                Set N             = #-of-entries
                High-level-SR8s   = 0
                Bottom-level-SR8s = 1

                Repeat

                  N = (N + IBC - 1) / IBC (truncate)

                  If N = 1, exit

                  If High-level-SR8s = 0,
                        High-level-SR8s    = 1
                        Bottom-level-SR8s  = N
                  Else High-level-SR8s = High-level-SR8s + N

          Set Total-SR8s = High-level-SR8s + Bottom-level-SR8s

Step 2:   Determine the number of SR8s per page:

                Calculate size of an SR8:

                        SR8-size = 32 + (IBC + 1) * (keylen + 8)

                Calculate number of SR8s per page:

                        (Page-size - Page-reserve - 32) / (SR8-size + 8)

Step 3:   Establish the INDEX DISPLACEMENT:

                If Number of Higher Level SR8s is less than 2, set
                the DISPLACEMENT = High-level-SR8s.  (For a one or
                two-level index, displacement will be 0 or 1
                respectively.)

                If Number of Higher Level SR8s is greater than 1,
                compute the displacement:

                  (High-level-SR8s + SR8s-per-page - 1)
                  -------------------------------------  + 1 (truncate)
                            SR8s-per-page

                If the calculate displacement is greater than the number
                of pages in the area containing the index, then:

                  Displacement = Number of pages in area / 2